UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

The top and beyond: missing energy and little Higgs in ATLAS

Rijpstra, M.

Link to publication

Citation for published version (APA):Rijpstra, M. (2011). The top and beyond: missing energy and little Higgs in ATLAS

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 03 Aug 2018

2The ATLAS Detector

This chapter describes the experimental setup that provides the data of interest: the ATLASdetector. It is one of the four particle detectors that are designed to measure physicsprocesses that occur at collisions of hadrons produced by the Large Hadron Collider. Inthe first section, the Large Hadron Collider is described. The subsequent sections cover theATLAS detector and each of its components.

2.1 The Large Hadron Collider

The Large Hadron Collider (LHC) is a circular particle accelerator situated at CERN in thetunnel where the Large Electron Positron accelerator (LEP) operated between the years1989 and 2000. The LHC is designed to collide beams of protons with 14 TeV center ofmass energy1, whereas LEP collided electrons on positrons with center of mass energies from91 GeV up to 209 GeV. The larger energy and the fact that the two beams of particles arenot mutual antiparticles require a more advanced accelerator machine.

The counter-rotating beams cross each other in four points along the tunnel, wherethe particle detectors ALICE [28], ATLAS, CMS [29] and LHCb [30] are located. Figure2.1 locates the detectors along the LHC ring as well as the pre-accelerators: the linearaccelerator (LINAC), the Proton Synchotron Booster (PSB), the Proton Synchotron (PS)and the Super Proton Synchotron (SPS). Protons are produced by stripping electrons fromhydrogen atoms and are subsequently accelerated to 50 MeV in the LINAC. The protonsare then accelerated in three steps by the circular pre-accelerators to 1 GeV, 26 GeV and450 GeV respectively. An arrangement of the protons into bunches of 1011 protons eachand a 25 ns bunch separation are established. The SPS finally injects the bunches both inclockwise and counter-clockwise direction into the LHC, which accelerates them to energiesup to 7 TeV.

The LHC [31] consists of eight arcs and eight straight sections, adding up to acircumference of 27 km. The two vacuum beam pipes are surrounded by several thousandsof superconducting magnets, which accomplish the bending and focusing of the beams. Asthe radius of the accelerator is fixed by the existing tunnel, the energy of the proton beams

1 Occasionally beams of lead ions are accelerated and collided at√

s =1148 TeV instead.

21

The ATLAS Detector

LHC

CMS

ALICE

LHCb

SPS

PS

PSBATLASPb−ions

protons

(not to scale)

LINAC

Figure 2.1: Schematic view of the particle accelerators and detectors at CERN.

is constrained by the strength of the bending magnets. The bending is achieved by 1232dipole magnets, which are cooled to 1.9 K by liquid helium and provide a field strength of8.33 T. A special twin aperture design of the magnets allows the two beams to be subjectto opposite magnetic fields while sharing the iron structure and helium vessel.

The actual acceleration of the proton beams is established in radio frequency cavities,where an oscillating electric field of 2 MV is generated. The oscillation frequency of 400 MHzis tuned to the bunch spacing of the beams.

In addition to the center of mass energy of the provided collisions, an importantcharacteristic of an accelerator machine is the beam intensity or luminosity L. It determinesthe rate of proton-proton interactions and thereby the rate of interesting events that occurin the center of the ATLAS detector:

dNevents

dt= Lσevents, (2.1)

where σevents is the cross section of those events at a given√s. The achieved luminosity of

an accelerator is described as

L =fNn2

A, (2.2)

where f is the revolution frequency, N is the number of bunches per beam, n is the numberof particles per bunch and A is the cross section of the beam. The target luminosity of theLHC, L = 1034 cm−2s−1, is reached by operating at the values given in Table 2.1.

The luminosity is subject to degradation due to beam losses from collisions andimperfections, resulting in an expected beam lifetime of approximately 15 hours [31].Refilling the LHC requires roughly 4 minutes per beam and ramping up the energy from450 GeV to 7 TeV in the LHC machine takes an additional 20 minutes, thus defining theminimum turnaround time. In a realistic scenario, the acquired integrated luminosity isestimated at 100 fb−1 per year when running at design luminosity.

Operating the LHC machine at design beam energies and design luminosity is a delicateprocedure for which no test setup exists. Initially, in 2008, a faulty electrical connection

22

2.2 The ATLAS Detector

Circumference 26659 mInjection energy 450 GeVBunch spacing 25 nsParticles per bunch n 1011

Bunches per beam N 2808Revolution frequency f 11 kHzBeam radius 16 µm

Table 2.1: Properties of the LHC at√s = 14 TeV proton-proton operation and design luminosity

L = 1034 cm−2s−1.

between two magnets caused severe damage to the accelerator and a full year was needed forrecovery. At the next attempt, in November 2009, the beams were maintained at injectionenergy to get acquainted with the steering and stabilizing. Thousands of events wererecorded with the ATLAS detector at

√s = 900 GeV and provided the first opportunity to

test reconstruction algorithms and detector performance with proton collisions2. The energywas gradually ramped up to reach the world record center of mass energy of 2.36 TeV onDecember 8th, 2009. The collision energies were expected not to exceed 10 TeV until 2012,after a scheduled maintenance shutdown, which is therefore the focus of the studies in thisthesis.

Ultimately, during summer 2010, a center of mass energy of 7 TeV was maintained anda total integrated luminosity of 3.46 pb−1 was recorded by ATLAS up until the sixth ofSeptember. The instantaneous luminosity reached a maximum of 1031 cm−2s−1, which is athousandth of the design luminosity. Figure 2.2 displays the course of integrated luminositydelivered by the LHC as well as the part recorded by the ATLAS detector. The recordedcollisions are investigated in Chapter 8.

2.2 The ATLAS Detector

The ATLAS detector was designed to observe the wide range of particles that are expectedto be created in proton-proton collisions at the unprecedented energies and luminosity.To this end, it is built up of several subdetectors, configured in concentric layers aroundthe interaction point, each optimized for the detection of a specific type of particles.Charged particles leave traces in the tracking detectors and all particles except for muonsand neutrinos deposit their energy in the calorimeters. Muons are detected by the muondetectors in the outermost layer, whereas neutrinos escape the detector without leaving atrace3.

From the interaction point outwards, the first subdetector of ATLAS is the innerdetector, one of the two tracking detectors. The subsequent subdetector is the calorimetersystem, divided into an electromagnetic and a hadronic component, which measures

2 During 2008 several runs of cosmic muon data were taken with ATLAS.3 Their transverse momenta can nevertheless be reconstructed by means of momentum conservation in

the transverse plane. This procedure is addressed in detail in Section 4.6 and Chapter 6.

23

The ATLAS Detector

Day in 2010

29/03 30/04 01/06 03/07 04/08 06/09

]-1

To

tal In

teg

rate

d L

um

ino

sity [

pb

0

1

2

3

4

5

Day in 2010

29/03 30/04 01/06 03/07 04/08 06/09

]-1

To

tal In

teg

rate

d L

um

ino

sity [

pb

0

1

2

3

4

5 = 7 TeVs ATLAS Online Luminosity

LHC Delivered

ATLAS Recorded

-1Total Delivered: 3.69 pb-1

Total Recorded: 3.46 pb

Figure 2.2: Integrated luminosity delivered by the LHC (blue) and recorded by ATLAS (yellow)up to September 6, 2010 during stable beams with

√s = 7 TeV. Figure taken

from [32].

the energy of particles by total absorption. The outer tracking detector is the muonspectrometer, designed to measure tracks of muons in particular. Both tracking detectorsoperate in a magnetic field, provided by a solenoidal and a toroidal magnet systemrespectively. Figure 2.3 illustrates the arrangement of the subdetectors and magnets thatcompose the ATLAS detector. In addition, the dimensions are shown, which indicate thatit is the largest particle detector ever built for an accelerator experiment. The detector wasinstalled in the underground cavern between 2003 and 2008 after many years of research,preparation and construction.

The Coordinate System

The origin of the ATLAS coordinate system is defined as the nominal interaction point inthe center of the detector. The z-axis runs parallel to the beam line in counterclockwisedirection. The half of the detector that corresponds to positive values of z is referred to asside A and the other half as side C. The x-axis points to the center of the LHC ring and they-axis points upwards to the surface, resulting in a righthanded orientation. The xy-planeis referred to as the transverse plane.

The ATLAS detector has a global cylindrical structure, where each subdetector consistsof concentric layers around the beam axis, the barrel component, and two endcaps formedby disks perpendicular to the z-axis on each side of the interaction point. A coordinatesystem closely related to cylindrical coordinates is convenient. The radial distance is givenby R =

√

x2 + y2. The azimuthal angle ϕ ∈ [−π, π] is the angle with the positive x-axisand increases in clockwise direction when looking down the positive z-axis. The polar angleθ ∈ [0, π] is defined as the angle with the positive z-axis, albeit generally replaced by the

24

2.2 The ATLAS Detector

Figure 2.3: Computer generated image of the ATLAS detector. A part is cut away in order toreveal the structure of concentric subdetectors that constitute the detector.

pseudorapidity η, which is given by

η = − ln

[

tan

(

θ

2

)]

. (2.3)

The preference for this quantity is motivated by the particle flux being roughly constant asa function of η. A direction (η, ϕ) is assigned to reconstructed final state objects and theopening angle between two of them is denoted ∆R, i.e.

∆R =√

(∆η)2 + (∆ϕ)2. (2.4)

Requirements

The performance requirements for the design of the ATLAS detector are based on theprocesses that may be observed at this new energy scale, such as the production of theHiggs boson, SUSY particles or heavy gauge bosons W ′ and Z ′. The extensive varietyof objects to be detected, the broad energy range of particles to be measured, thehigh radiation conditions and the high collision rate impose strict requirements on thedetector’s precision, speed, performance, radiation hardness, efficiency and acceptance.The performance requirements in terms of resolution as well as the acceptance of eachsubdetector are summarized in Table 2.2. An additional challenge is the instantaneous

25

The ATLAS Detector

selection of collisions to be stored, which is taken care of by the trigger system. This isaddressed in Section 2.2.5, after the three subdetectors and the magnet system are describedin more detail in Sections 2.2.1 through 2.2.4.

Subdetector Required Resolution η coverage

Inner Detector σ(pT)/pT = 0.05% pT ⊕ 1% |η| < 2.5

Electromagnetic Calorimeter σ(E)/E = 10%/√

E(GeV) ⊕ 0.7% |η| < 3.2

Hadronic Calorimeter σ(E)/E = 50%/√

E(GeV) ⊕ 3% |η| < 3.2

σ(E)/E = 100%/√

E(GeV) ⊕ 10% 3.1 < |η| < 4.9

Muon Spectrometer σ(pT)/pT = 10% at pT = 1 TeV |η| < 2.7

Table 2.2: Performance requirements for the subdetectors of the ATLAS detector [33].

2.2.1 The Inner Detector

The inner detector is the subdetector closest to the interaction point, where the density ofparticles is largest. High granularity and good radiation tolerance are required. With aninner radius of 45 mm it is as close as 10 mm to the beam pipe and it extends to a radiusof 1150 mm. It is contained inside a solenoidal magnetic field of 2 T (cf. Section 2.2.4)and designated to the reconstruction of tracks of charged particles. In order to achievehigh momentum resolution over the entire range of momenta as well as precise vertexmeasurements, the inner detector consists of three complementary tracking devices: the pixeldetector, the semi-conductor tracker (SCT) and the transition radiation tracker (TRT).Their configuration is illustrated in Figure 2.4.

The Pixel Detector

The innermost part of the inner detector is the pixel detector, consisting of three concentriclayers around the beam axis and three discs perpendicular to the beam axis on each side ofthe interaction point. The 1744 modules that constitute the pixel detector each consist ofa 250 µm layer of silicon implanted with readout pixels that measure 50× 400 µm2, addingup to a total of 80 million pixels. Electron-hole pairs are created in the silicon when chargedparticles pass through and a current is induced due to the p-n junction in the doped silicon,which is read out by the pixels. The pixel detector primarily contributes to the preciseidentification of the primary vertex and secondary vertices.

The Semi-Conductor Tracker

The middle component of the inner detector is the SCT, which is composed of fourconcentric barrels around the beam axis and nine endcap disks along the beam line on eachside. It thus extends radially from 255 mm to 549 mm and longitudinally from 810 mm to2797 mm from the interaction point. Its detection principle is similar to that of the pixel

26

2.2 The ATLAS Detector

Figure 2.4: The inner detector and the configuration of its three components: the pixel detector,the SCT and the TRT.

detector, although the lower particle density allows for long, narrow silicon strips ratherthan small rectangular pixels. The strips are configured in two layers under a small anglewith respect to each other, such that a position measurement along the strip length can beobtained from hits in overlapping strips.

The Transistion Radiation Tracker

The technology employed in the outer component of the inner detector, the TRT, is twofold:gaseous straw tubes are interleaved with transition radiation material. The barrel contains73 such layers and the 20 wheels on each endcap are covered with 160 such planes, addingup to approximately 350k readout channels.

When charged particles pass through, the gas inside the tubes is ionized and a voltagedifference between the tube and the anode wire in its center causes the free electrons todrift towards the wire. The drift time is converted into the distance of the track to the wire.

Transition radiation is emitted when highly relativistic charged particles pass thetransition between two materials with different dielectric constants. The intensity of thetransition radiation photons is proportional to the Lorentz factor of the traversing particle,which is much higher for electrons than for pions, at equivalent energies, due to theirmass difference. The gas mixture inside the straw tubes contains xenon, which absorbs theradiation photons and thus produces a signal with a high amplitude when an electron passesthrough.

The readout electronics of the tubes apply two distinct thresholds: a lower one that

27

The ATLAS Detector

detects the ionization clusters and a higher one that is optimized for transition radiationfrom electrons and allows for rejection of tracks from π± background.

A particle originating from the interaction point, given that it satisfies |η| < 2.0 andpT > 5.0 GeV, typically gives rise to three pixel hits, four SCT measurements andaround 30 TRT hits. The semiconductor trackers provide three-dimensional space pointswith high precision, yet the high number of TRT hits over the larger part of the tracklength contributes significantly to the momentum measurement. In addition, efficientelectron identification is provided by the TRT. The expected inverse transverse momentumresolution on a reconstructed track with 0.25 < |η| < 0.50 is [33]

σ(1/pT) = 0.34 TeV−1(1 ⊕ 44 GeV/pT), (2.5)

where the first term indicates the expected asymptotic resolution at infinite transversemomentum and the second term represents the impact of multiple scattering at lowmomenta. The expected resolution on the transverse impact parameter is

σ(d0) = 10 µm(1 ⊕ 14 GeV/pT). (2.6)

2.2.2 The Calorimeters

After having traversed the inner detector, particles enter the calorimeter system, which issituated outside the solenoidal magnet that surrounds the inner detector. It extends fromapproximately 1.4 m to 4.2 m from the interaction point in the transverse plane. Firstlyencountered is the electromagnetic calorimeter, which is optimized for the identificationand energy determination of photons and electrons. The hadronic calorimeter is dedicatedto the reconstruction of hadronic showers from quarks, gluons and hadronically decayingtaus. Altogether, the calorimeter system covers the full azimuth and the pseudorapidityrange |η| < 4.9. The configuration of the calorimeters is depicted in Figure 2.5.

Muons generally deposit a mere fraction of their energy in the calorimeters and continueto be detected by the muon spectrometer (cf. Section 2.2.3). Neutrinos remain undetectedentirely. The transverse component of the undetected energy can be nevertheless estimatedby means of the expected energy balance in the transverse plane. The performance of thecalorimeters is of direct influence on this quantity, the missing transverse energy, which willbe discussed in more detail in Chapter 6.

Both the electromagnetic and the hadronic calorimeter consist of sampling detectors,i.e. layers of passive, dense material alternated with layers of active material. The passivematerial causes incident particles to initiate a shower or cascade of secondary particles,which are detected in the active material. In sufficient successive layers, the primary particlewill have transferred all its initial energy.

Electromagnetic showers are the result of Bremsstrahlung and e+e− pair production andthe characteristic interaction distance is the radiation length X0 of the material4. Hadronicshowers are the result of nuclear interactions and develop over larger distances. The required

4 X0 is the mean distance over which an electron loses all but 1/e of its energy.

28

2.2 The ATLAS Detector

depth of the material for complete containment of the shower is larger and is expressed interms of the nuclear interaction length λ of the passive material.

Figure 2.5: Configuration of the ATLAS calorimeters, indicating the electromagnetic compo-nents as well as the tile calorimeters.

The Electromagnetic Calorimeter

The electromagnetic calorimeter consists of a barrel that covers |η| < 1.475 and two endcapwheels at 1.375 < |η| < 3.200 The passive material employed in the electromagneticcalorimeter are lead plates folded into an accordion shape. The space between the platescontains a honeycomb structure that is filled with liquid argon. Charged particles producedin showers induce free charge by ionizing the liquid argon, which is collected on the readoutelectrodes. The barrel component shares its cryostat vessel with the solenoid magnet (cf.Section 2.2.4) in order to minimize the amount of inactive material. Between the barrel andeach endcap wheel, around |η| = 1.4, some space is available for cables and services for theinner detector. The thickness of the electromagnetic calorimeter varies from 22X0 to 33X0.

The modules of which the electromagnetic calorimeter is composed are divided into threelongitudinal layers, as illustrated in Figure 2.6. The front layer is finely segmented in η,which facilitates γ/π0 separation. The middle layer is thickest and receives the larger part ofthe energy deposited by electromagnetic showers. The third layer has a coarse granularityand is mainly used to recover the tails of highly energetic electromagnetic showers andto discriminate between hadronic and electromagnetic showers based on the larger energydeposit by the former.

29

The ATLAS Detector

The achieved resolution on the energy E as measured in a test beam of electrons [34] is

σ(E)

E=

10%√

E(GeV)⊕ 0.17%. (2.7)

The first term represents the stochastic response of the calorimeter, whereas the constantterm is due to systematic effects, such as non-uniformity and stability.

∆ϕ = 0.0245

∆η = 0.02537.5mm/8 = 4.69 mm ∆η = 0.0031

∆ϕ=0.0245x4 36.8mmx4 =147.3mm

Trigger Tower

TriggerTower∆ϕ = 0.0982

∆η = 0.1

16X0

4.3X0

2X0

1500

mm

470

mm

η

ϕ

η = 0

Strip cells in Layer 1

Square cells in Layer 2

1.7X0

Cells in Layer 3 ∆ϕ×�∆η = 0.0245×�0.05

Figure 2.6: Schematic view of a module in the electromagnetic calorimeter, showing the typicalaccordian shape and the granularity of the different layers.

The Hadronic Calorimeter

The hadronic calorimeter surrounds the electromagnetic calorimeter and constitutes ascintillator tile calorimeter at |η| < 1.7 and two endcap wheels at 1.5 < |η| < 3.2. Thetile calorimeter in turn is divided into a central barrel at |η| < 1 and two extended barrelsat 0.8 < |η| < 1.7. The gap in between contains cables, services and power supplies forthe inner detector as well as for the electromagnetic calorimeter. The passive material issteel, which functions simultaneously as return yoke for the solenoid magnet. The activemedium is formed by scintillating plastic tiles that emit the absorbed energy in the formof light. The scintillation light is picked up by wavelength shifting fibers and propagatedto photomultiplier tubes, where the signal is amplified and detected. The thickness of thehadronic calorimeter is approximately 10 λ. It follows from pion test beam results [34] thatthe achieved energy resolution meets the requirement

σ(E)

E=

50%√E(GeV)

⊕ 3%. (2.8)

30

2.2 The ATLAS Detector

Hadronic showers are more complex than electromagnetic showers and the resolution islimited by binding energy losses and non-compensation, i.e. the electromagnetic componentof the shower is detected more efficiently than the hadronic component. The electromagneticfraction is energy dependent and its fluctations contribute to the resolution. The ratio ofthe electron response and pion response is determined as a function of energy in test beamsand used in the calibration (cf. Section 4.2).

In the hadronic endcap wheels the passive layers are made of copper and the activemedium in between is liquid argon. The readout cells measure ∆η ×∆ϕ = 0.1× 0.1 in theregion 1.5 < |η| < 2.5 and 0.2 × 0.2 in the more forward region.

The Forward Calorimeter

In order to estimate the missing transverse energy, as large hermetic calorimeter coverage aspossible is pursued. The coverage in the very forward region, 3.1 < |η| < 4.9, is provided bythree wheels on either side: one electromagnetic component and two hadronic components.With inner radii of approximately 8 cm, they are situated close to the beam and the expectedradiation level is high. Closest to the interaction point is the electromagnetic component inwhich copper acts as the passive material. The two hadronic components employ tungstenand the active material in all three of them is liquid argon. On each side, the forwardcalorimeter wheels share the liquid argon cryostat with the electromagnetic and hadronicendcaps.

2.2.3 The Muon Spectrometer

The muon spectrometer is the largest and outermost subdetector of ATLAS. With inner andouter radii of approximately 4.5 m and 11 m respectively and stretching out from about 7 mto 23 m from the interaction point on each side in the longitudinal direction, it occupies avolume of around 16000 m3. It was designed to trigger on muons with high momenta, whichplay a role as a distinguishing feature in several interesting physics channels, as well as toreconstruct the tracks of muons that pass through with high precision. The componentsproviding the first functionality are the Resistive Plate Chambers (RPC) and the ThinGap Chambers (TGC), while the latter is achieved by the Monitored Drift Tube (MDT)chambers and the Cathode Strip Chambers (CSC). A three-dimensional representation ofthe muon spectrometer is shown in Figure 2.7(a), indicating the four different types ofcomponents. The arrangement is such that a particle originating from the interaction pointwill traverse three layers of muon stations as it is bended by the magnetic field (cf. Section2.2.4).

The Barrel

The barrel of the muon spectrometer consists of three concentric cylindrical layers of muonstations that cover space up to a pseudorapidity of |η|=1 and in nearly full azimuth. Thestations are organized in sixteen sectors, alternating small (S) and large (L), thus followingthe structure of the eight barrel toroid magnet coils (cf. section 2.2.4). Muon stations in theinnermost layer are single MDT chambers located just outside the hadronic calorimeter andnamed Barrel Inner (BI) chambers. Stations in the middle layer consist of a Barrel Middle

31

The ATLAS Detector

ChambersChambers

Chambers

Chambers

Cathode StripResistive Plate

Thin Gap

Monitored Drift Tube

(a) A three-dimensional schematic view of the muonspectrometer.

+x

+y

02

03

04 05 06

07

08

09

10

11

121314

15

16

01

(b) A schematic view of the transverse crosssection of the muon spectrometer, indicating thenumbering convention of the sectors as well asthe arrangement of the muon stations around thebarrel toroid magnet.

Figure 2.7: The muon spectrometer.

(BM) MDT chamber with a RPC on either side and are situated inside the barrel toroidmagnet. The outer layer consists of stations that each comprise a Barrel Outer (BO) MDTchamber and a RPC and are positioned just outside the barrel toroid magnet. A transversecross section of the barrel of the muon spectrometer is shown in Figure 2.7(b), indicatingthe numbering scheme of the sixteen sectors.

The chamber coverage is limited in the region |η| < 0.1 to accommodate inner detectorand calorimeter services as well as in the barrel/endcap transition region where most ofthe middle stations are not installed yet for first operation. In sectors 12 and 14, aroundϕ ∼ − 9

24π and ϕ ∼ −15

24π, the support feet of the toroidal magnet system prevent full

azimuthal coverage.

The Endcaps

The two endcaps of the muon spectrometer consist of four disks each and cover apseudorapidity range of 1.0 < |η| < 2.7. The greater part of the disks consists oftrapezoidally shaped MDT chambers, yet the innermost disk of the innermost layer of eachendcap is equipped with CSCs. Thus, in the pseudorapidity range 2.0 < |η| < 2.7 in whichthe largest particle flux is expected, a better spatial resolution and faster measurements areaccomplished. The trigger information in the forward regions is provided by three planesof TGCs in each endcap, which cover 1.05 < |η| < 2.4.

32

2.2 The ATLAS Detector

Precision Chambers

The precision measurement of muon trajectories is accomplished by MDTs and CSCs,designed to reach a transverse momentum resolution of approximately 10% for muonmomenta of 1 TeV.

MDTs consist of two multilayers of aluminium tubes and a support structure, as shownin Figure 2.8. The cathode tubes of 30 mm diameter are filled with a gas mixture and

Multilayer

Spacer frame

Figure 2.8: Schematic layout of a monitored drift tube chamber.

contain an anode wire to which a voltage is applied during operation. When a chargedparticle passes through, ionization clusters are created in the gas, which will drift to theanode wire where the signal is propagated to the readout electronics. The position of theincident particle is derived from the timestamp corresponding to the signal pulse passing athreshold of five times the noise level by making use of the relation between the drift timeand the distance of the particle to the wire.

CSCs consist of four planes of anode wires and two cathode planes equipped with strips.The coordinates of a traversing charged particle are obtained from the relative measurementof induced charge on adjacent cathode strips. The strips on each of the two cathode planesare positioned orthogonally, thus allowing for determination of two coordinates.

The actual reconstruction of muon tracks is further described in Section 4.5.

Trigger Chambers

Trigger chambers serve the purpose of providing rapid information about charged particlestraversing the muon system.

RPCs are gaseous detectors with a time resolution of 1.5 ns and a typical spatialresolution of 1 cm. They consist of two rectangular detectors, contiguous to each other,each composed of two gas volumes and two readout strip panels. A gas volume is enclosed

33

The ATLAS Detector

by two resistive plates separated by 2 mm to which a voltage is applied, such that avalanchesinduced by ionizing tracks are accelerated towards the anode plane, where the signal is readout.

TGCs are based on the same technology as CSCs, although the spacing between thewires and the cathodes is relatively small in order to guarantee a drift time that does notexceed the 25 ns LHC bunch separation.

2.2.4 The Magnet System

The ATLAS magnet system generates a magnetic field configuration such that thetrajectories of charged particles are bended when traversing the tracking devices, the innerdetector and the muon spectrometer. It consists of two superconducting magnet systems, atoroidal system and a central solenoid, that add up to a diameter of 22 m and a length of26 m. The toroidal magnet system provides a magnetic field inside the volume of the muonspectrometer, while the solenoidal magnet generates a homogeneous field parallel to thebeam axis inside the inner detector. The curvature of the trajectory followed by a chargedparticle when passing through the field is used to determine its momentum. A schematicview of the configuration of the toroidal and solenoidal magnets in ATLAS is depicted inFigure 2.9.

Figure 2.9: Three dimensional schematic view of the magnet system, showing the eightrectangular barrel coils, eight endcap coils on either side and the solenoid in thecenter.

The Toroidal Magnet System

The toroidal magnet system is built up of a barrel toroid and two endcap toroids.

The barrel toroid consists of eight superconducting rectangular coils, each encased ina cryostat. The total assembly weighs 830 tons and adds up to 25.3 m axial length andinner and outer diameters of 9.4 m and 20.1 m respectively. Cooling down to the nominal

34

2.2 The ATLAS Detector

operational temperature of 4.6 K takes 5 weeks. The field strength provided by the barreltoroid at the nominal operational current of 20.5 kA varies from 0.15 T to 2.5 T.

The endcap toroid systems consist of eight coils each, which are located interleaved withthe barrel toroid coils on either side, thus generating a magnetic field in the endcap regionsof the muon spectrometer. With an inner and outer diameter of 1.65 m and 10.7 m andan axial length of 5.0 m each endcap toroid weighs 239 tons. Powered in series with thebarrel toroid, the endcap toroids generate a field strength that varies from 0.2 T to 0.35 Tat nominal operational current.

The Central Solenoid

The solenoidal magnet system is aligned with the beam axis and produces an axial fieldthroughout the volume of the inner detector. At the 7.730 kA nominal operational current,the strength of the field varies from 2 T at the interaction point to 0.9 T. With an axiallength of 5.8 m and a diameter of about 2.5 m, it is embedded inside the electromagneticcalorimeter. In contemplation of a minimal amount of material in front of the calorimeters,the solenoid shares its cryostat with the electromagnetic calorimeter.

In order to monitor the magnetic field inside the inner detector, four NMR (NuclearMagnetic Resonance) probes are mounted on the wall of the vessel at z=0, equally spacedin ϕ, that measure the magnitude of the magnetic field, |B|, with an accuracy of 10 µT.

Magnetic Field Sensors in the Muon Spectrometer

A precise knowledge of the strength of the magnetic field, B, at each given point insidethe volume of the muon spectrometer is required to reconstruct a muon’s momentum. Tothis end a magnetic field map is computed from contributions to the Biot-Savart integralfrom the toroidal and solenoidal magnet systems as well as other ferromagnetic materialpresent in the various detector systems. During operation however, the exact position ofthese components is continuously influenced by temperature fluctuations and material-induced magnetic forces. Therefore approximately 1800 magnetic field sensors are installedthroughout the MDT system to support the calculation of the magnetic field. Their positionin ATLAS is displayed in Figure 2.12(c). Each sensor is equipped with Hall probes on threeof its orthogonal faces that determine the components of the magnetic field from the inducedvoltage due to the Hall effect. The number of sensors mounted on a MDT chamber variesfrom zero to four, each including its own readout electronics and a temperature sensor toallow for local calibration. By measuring the response of each Hall probe as a function offield strength, field orientation and temperature T in test stands at CERN [35], the Hallvoltage of each sensor was calibrated. The achieved accuracies on the magnitude |B| of themagnetic field are 0.2 mT up to |B|=1.4 T and 1 mT up to 2.5 T. In addition, two NMRprobes are mounted on the barrel toroid system in order to monitor potential long-termdrifts in the response of the Hall probes.

A number of criteria are defined in order to assess the quality of the value read out byeach magnetic field sensor. The criteria include, amongst other,

◦ |T − 20◦C| < 10◦C;

◦ |B| < 2.0 T in the endcaps and |B| < 3.0 T in the barrel;

35

The ATLAS Detector

◦ RMS(|B|/I) < 1.0 T/A;

◦ RMS(T ) < 2◦C;

◦ |Bi − 〈Bi〉 | < 3 σ(Bi) for each component i ∈ {x, y, z}, where the average is takenover all sensors;

◦ |T − 〈T 〉 | < 3 σ(T ) where the average is taken over all sensors.

In case any of the criteria is not satisfied, the sensor is flagged and disregarded in the fieldmap calculation.

Magnet Operation

After installation in ATLAS, the barrel toroid, each of the endcap toroids and the solenoidwere cooled down to 4.6 K and ramped to their nominal current during separate standalonetests. In 2008, the barrel toroid and endcap toroids were powered in series and commissionedas an assembly. During the first duration test of the combined configuration, the currentwas initially ramped up to 20400 A in about three hours time, subsequently ramped up toslightly over the 20500 A nominal operational current and finally the magnets were slowlydischarged. The behaviour of the toroid current and the magnetic field sensors mounted onMDT chambers of type BOL (Barrel Outer Large) is shown in Figures 2.10(a) and 2.10(b)respectively.

(a) The toroid current versus time during the firstduration test of the toroidal magnet system.

(b) The magnitude of the magnetic field asmeasured by the magnetic field sensors on MDTchambers of the type BOL versus time during thefirst duration test of the toroidal magnet system.

Figure 2.10: The first duration test of the toroidal magnet system.

Figure 2.11 displays the RMS of the magnitude of the magnetic field as read out by themagnetic field sensors during the current plateau of 20504.8 A that was part of the duration

36

2.2 The ATLAS Detector

test. The figure contains the measurements of all 95% of the magnetic field sensors thatpassed the quality criteria at that moment. In addition, the distribution is displayed formeasurements taken in 2010 during full operation. The number of sensors passing all criteriahas increased to 100% and the RMS values of the measurements have not degraded in twoyears time. The relative RMS values with respect to the mean value of |B| are of the orderO(10−5), which meets the requirement 5 · 10−4 from muon reconstruction performance [33].

Figure 2.11: The RMS of the magnitudeof the magnetic field duringconstant current as read outby the magnetic field sen-sors throughout the MDTsystem.

(mT)BRMS 0 0.005 0.01 0.015 0.02

num

ber

of s

enso

rs

0

50

100

150

200

250

300

350 2008

2010

Figure 2.12(c) shows the magnitude |B| of the magnetic field measured by each sensorduring operation, i.e. at 7.730 kA solenoid current and 20.5 kA toroid current.

The temperature fluctuations of the four sensors on a MDT chamber in the topsector, BIL2C05, are displayed in Figure 2.13. The temperature decreases slightly duringthe night and starts incrementing again around 7:00. It turns out that the temperaturedifference between the sensor locations on a chamber is considerable. The fluctuations ofthe temperature are accounted for in the calibration of the magnetic field components beingread out by the Hall probes. The RMS on the temperature measurement is well within theallowed range not to be flagged for each of the sensors in the figure.

2.2.5 The Trigger System

Operating at a bunch crossing frequency of 40 MHz, with approximately 23 interactionsoccurring per bunch crossing, the LHC will produce as many as one billion events per second.Merely a fraction of these events are of interesting nature as most of the interactions are socalled minimum bias events, i.e. partonic interactions with transverse momenta too smallfor perturbation theory to be valid. Moreover, it is not feasible to store the correspondingamount of data, approximately 1 PB s−1, on technical grounds. Therefore, a highly efficientselection of interesting events within a minimal time span is required. To this end, a complextrigger system is developed that reduces the rate of events to be stored by a factor of O(107).The event rejection procedure takes place in three subsequent stages: the Level-1 Trigger,the Level-2 Trigger and the Event Filter.

37

The ATLAS Detector

x (cm)

-100 -80 -60 -40 -20 0 20 40 60 80 100

y (c

m)

-100

-80

-60

-40

-20

0

20

40

60

80

100

|B| (

T)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

(a) The projection onto the transverse plane.

z (cm)

-150 -100 -50 0 50 100 150

y (c

m)

-100

-80

-60

-40

-20

0

20

40

60

80

100

|B| (

T)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

(b) The longitudinal projection onto the yz-plane

z (cm)

-150 -100-50 0 50

100 150x (cm)

-100

-50

0

50

100

y (c

m)

-100

-50

0

50

100

(c) Three dimensional view.

Figure 2.12: The magnitude of the magnetic field during operation as read out by sensorsthroughout the MDT system.

The Level-1 Trigger

The first level of the trigger system bases its decision on coarse granularity informationfrom the calorimeters and information from the muon trigger chambers only. The decisiontime is constrained to 2 µs by the pipeline memory in which all detector channel data of abunch crossing are stored until a decision is reached. In order to minimize the propagationtime through cables, the dedicated electronics are located as close as possible to the ATLASdetector. The Level-1 Trigger defines Regions of Interest (RoIs) in (η, ϕ) space where object

38

2.2 The ATLAS Detector

Figure 2.13: The temperate measuredby the four sensors onthe MDT chamber BIL2C05during part of the durationtest.

Time and Date17/0814:00

17/0820:00

18/0802:00

18/0808:00

18/0814:00

C)

°T

empe

ratu

re (

20

21

22

23

24

25

26

BIL2C05

sensor 1

sensor 3

sensor 2

sensor 0

candidates satisfy certain energy thresholds. Based on the multiplicity of RoIs, an event ispassed on to the Level-2 Trigger or rejected, which results in a reduction of the event rateto 75 kHz.

The Level-2 Trigger

The second level of the trigger system refines the Level-1 Trigger decision by using fullgranularity information from all detectors, including the inner detector. Dedicated softwareexamines the previously defined RoIs in more detail and attempts to reconstruct physicsobjects, i.e. electrons, photons, muons and jets, and it calculates the missing transverseenergy from the information from the calorimeters. Subsequently, a set of selection criteriais applied and the resulting event rate is 2 kHz.

The Event Filter

The Event Filter is designed to reduce the event rate from 2 kHz to the 200 Hz permanentstorage acceptance. In addition, it organizes the data into streams based on physics objectcandidates, as further described in Section 4.1. It accesses the full precision data from thecomplete detector and generally makes use of the same sophisticated algorithms as thereconstruction of the events that are eventually stored (cf. Chapter 4), yet looser criteriaare applied. The available processing time per event is approximately 4 s.

The sequence of algorithms that define a certain object candidate at each stage will bereferred to as a trigger chain. The final energy threshold and quality requirements arenaturally determined by the last stage, the Event Filter. The naming convention for atrigger chain is [N][TYPE][THRESHOLD](i) [QUALITY], where TYPE specifies the objectcandidate (as indicated in Table 2.3), N indicates its multiplicity, THRESHOLD is a numbercorresponding to a (transverse) momentum threshold, i indicates isolation and QUALITY

39

The ATLAS Detector

refers to the severity of requirements in the algorithm (e.g. loose, medium or tight in caseof electrons, cf. Section 4.4).

Object TYPE

muon muelectron ephoton g

electron/photon emjet j

forward jet fjtau tau

total energy tetotal jet energy tjmissing energy xeminimum bias mb

Table 2.3: The naming convention for TYPE in trigger chains.

The trigger chain e20 loose for instance, which is used in the following chapters, isdesigned to trigger on electrons. The algorithms in this chain aim to be as efficient aspossible for electrons with a transverse momentum larger than 20 GeV and which satisfythe loose requirements defined in the reconstruction algorithm.

The available trigger chains are defined in terms of a trigger menu, which varies in timewith increasing instantaneous luminosity. The events that pass any trigger chain in the givenmenu are arranged in luminosity blocks, each containing typically a couple of minutes of datataking, and stored. The beam conditions and detector performance are stored per luminosityblock. Subsequently, the collected data is passed on to the reconstruction software, whichis described in Chapter 4.

40